Editorial on censoring in prospective studies by Jungmalm et al.
Jungmalm et al. have recently published an educational editorial in BJSM concerning censoring in prospective studies.
The editorial clarifies the importance of censoring when estimating cumulative incidence proportion (CIP). Its focus is sport injuries.
In Cross-sectional studies, the prevalence proportion is the number of injuries divided with the total sample size. By employing this approach to cohort studies, it is assumed that all participants are at risk throughout the entire follow-up period. In other words, it is assumed that none of the runners quits because of other reasons than injury, and that all injuries occur at the last day of the follow-up period. This approach will often lead to underestimation of CIP, because these assumptions are seldom realistic in prospective studies in a sports medicine setting.
“The editorial involves an example from ProjectRun21, where the censoring partly affects the CIP to be twice as high as the CIP without censoring. Specifically, the CIP in the study ProjectRun21 is 22% without censoring compared to 44% with censoring” Rasmus Nielsen, one of the co-authors of the educational editorial, explains. According to the authors, researchers in Sports medicine are encouraged to use censoring when estimating CIP in prospective studies.